Real time control of an adaptive vehicular occupant restraint system using intelligent control techniques

Thousands of people die every year in car accidents. Existing restraint systems lack the capability of varying their output during the impact event. This research applies intelligent control techniques in the design of a real time controller for an adaptive restraint system to dramatically minimize occupant injuries during frontal crashes. The controller first task is to predict the maximum desired head and chest acceleration levels in a particular crash event. A methodology to develop a predictive Adaptive Neuro-Fuzzy Inference System (ANFIS) model to accomplish this aim is introduced. The controller's second task is to keep adjusting the restraint system continuously in order to maintain the actual head and chest accelerations below the targeted thresholds. Different configurations of fuzzy logic controllers FLCs are also designed to accomplish this aim. To optimize the FLCs, the controllers' parameters needed to be tuned. Tuning based on trial-and-error turned to be a tedious and time-consuming task and therefore a methodology to simplify the automatic tuning process of self organized FLCs is introduced. Due to the prohibitive cost of crash testing our system was designed and tested first with a lumped parameter model, second with a multi-body model and finally by using a complete crash simulation environment that link Madymo and Matlab.